vane
StarGet Started →
Vane Data / Contributing

Development

This page summarizes the project development flow.

Repository setup

shell
git clone --recursive https://github.com/AstroVela/vane.git
cd vane

If needed:

shell
git submodule update --init --recursive

Build dependencies

The project uses scikit-build-core and CMake.

Python build dependencies include:

  • scikit-build-core>=0.11.4
  • pybind11[global]>=2.6.0
  • cmake>=3.29.0
  • ninja>=1.10

On Debian/Ubuntu:

shell
sudo apt-get update
sudo apt-get install -y build-essential cmake ninja-build pkg-config curl zip unzip tar flex bison

Editable build

shell
python -m pip install -e . --no-build-isolation -v

With vcpkg:

shell
git clone https://github.com/microsoft/vcpkg.git ../vcpkg
../vcpkg/bootstrap-vcpkg.sh


python -m pip install -e . --no-build-isolation -v \
  --config-settings=cmake.define.CMAKE_TOOLCHAIN_FILE="$(pwd)/../vcpkg/scripts/buildsystems/vcpkg.cmake"

Formatting

The source CONTRIBUTING.md describes formatting through scripts/format.

Examples:

shell
scripts/format root --changed
scripts/format submodule --changed
scripts/format workspace --changed

The root formatter avoids scanning external/duckdb by default. Use the submodule command when changing DuckDB internals.

Tests

CI builds and installs Vane Data on Ubuntu 24.04 with Python 3.12, then runs:

shell
python -m pytest tests/fast

For a focused local test:

shell
python -m pytest tests/fast/test_udf_process.py

Source areas

  • vane/: Vane Data public Python API and AI helpers.
  • duckdb/execution/: Python UDF executor routing and runtimes.
  • duckdb/runners/: native and Ray runner code.
  • src/duckdb_py/: C++ Python bindings.
  • external/duckdb/: DuckDB submodule.
  • examples/: runnable script examples.
  • Benchmark workloads: performance and fault-tolerance experiments.

Contribution rule of thumb

Keep changes scoped. If a change affects public APIs such as map_batches, vane.configure, or AI helpers, update docs and tests in the same change.

← Sizing